Re: [Haskell-cafe] Increasing Haskell modularity
It seems to me the far easier choice, in ditching the feature “not worth the problem” is simply to eliminate orphan instances as in “Extension 1” described in Edward Yang’s blog post (http://blog.ezyang.com/2014/09/open-type-families-are-not-modular/). I’d like to open up a further discussion on this at some future point, as I’m convinced that there is no “legitimate” use of orphans in the Haskell ecosystem that cannot be substituted for with other safer and more straightforward techniques. -g On October 1, 2014 at 9:26:45 AM, Gesh (gesh@gesh.uni.cx) wrote:
Dear Café, Edward Yang recently wrote about the non-modularity of the current semantics for typeclasses[0] and open type families[1]. Given these problems, we can either try to work around the problem, or decide whether these features are worth the problems. The specific pair of proposals I am making are the following: * Give up on global uniqueness of typeclass instances. Oleg discusses adding local typclass instances without losing coherence in his reflection paper[2]. If we adopt his proposal, and weaken global uniqueness of instances to local uniqueness, we can then say that typeclass instances in a module are considered more local than imported instances, and therefore override them. Since instances will be locally unique, we don't have to pass around dictionaries. Data types that expect globally unique instances, such as Set, will have to store dictionaries, however. * Give up on non-associated open type families. This I have researched less well, but it seems to me that these have no use case. Since one can make whatever instances one wishes for these type families, one can't do anything useful with values of their instances. In other words, values of type F a are morally of type (forall a. a). Thus, the only way I can see these being useful is if anyone creating an instance of the family must also provide functions that can work with these instances, which brings us back to the case of associated types.
As far as I can see, these two features are useless and impeding the drive to make Haskell more modular. Therefore, my questions for you are twofold: * Is there a reason to keep these features? * If not, why don't we drop them?
Awaiting your criticism and ire, Gesh
[0] - http://blog.ezyang.com/2014/07/type-classes-confluence-coherence-global-uniq... [1] - http://blog.ezyang.com/2014/09/open-type-families-are-not-modular/ [2] - http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.112.7819 _______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Gershom, As far as Haskell is concerned, yes, I think orphan instances should go. However, they are very useful! It is our aversion to using them which prevents us from splitting things into separate, modular libraries: for instance, library A provides type classes, library B provides unrelated types, library C provides A instances for B types. When people use the word "modularity", it is assumed that this is one of the things it can do, so it's kind of funny to hear the word used in a sense which rules out this case. But as you and Ed have argued in the past, orphan instances in Haskell actually cause a fair amount of problems, which I consider to be a *ringing* indictment of type classes as the only means of modular abstraction. If we outlaw the orphans, we'll rule out some bad cases, and at the very least make crystal clear the actual problem of modularity such that people will not easily confuse Haskell type classes with a solution to it, as some do now. So I'm in favor. Kind regards, Jon On Wed, Oct 1, 2014, at 08:25 AM, Gershom B wrote:
It seems to me the far easier choice, in ditching the feature “not worth the problem” is simply to eliminate orphan instances as in “Extension 1” described in Edward Yang’s blog post (http://blog.ezyang.com/2014/09/open-type-families-are-not-modular/).
I’d like to open up a further discussion on this at some future point, as I’m convinced that there is no “legitimate” use of orphans in the Haskell ecosystem that cannot be substituted for with other safer and more straightforward techniques.
-g
On October 1, 2014 at 9:26:45 AM, Gesh (gesh@gesh.uni.cx) wrote:
Dear Café, Edward Yang recently wrote about the non-modularity of the current semantics for typeclasses[0] and open type families[1]. Given these problems, we can either try to work around the problem, or decide whether these features are worth the problems. The specific pair of proposals I am making are the following: * Give up on global uniqueness of typeclass instances. Oleg discusses adding local typclass instances without losing coherence in his reflection paper[2]. If we adopt his proposal, and weaken global uniqueness of instances to local uniqueness, we can then say that typeclass instances in a module are considered more local than imported instances, and therefore override them. Since instances will be locally unique, we don't have to pass around dictionaries. Data types that expect globally unique instances, such as Set, will have to store dictionaries, however. * Give up on non-associated open type families. This I have researched less well, but it seems to me that these have no use case. Since one can make whatever instances one wishes for these type families, one can't do anything useful with values of their instances. In other words, values of type F a are morally of type (forall a. a). Thus, the only way I can see these being useful is if anyone creating an instance of the family must also provide functions that can work with these instances, which brings us back to the case of associated types.
As far as I can see, these two features are useless and impeding the drive to make Haskell more modular. Therefore, my questions for you are twofold: * Is there a reason to keep these features? * If not, why don't we drop them?
Awaiting your criticism and ire, Gesh
[0] - http://blog.ezyang.com/2014/07/type-classes-confluence-coherence-global-uniq... [1] - http://blog.ezyang.com/2014/09/open-type-families-are-not-modular/ [2] - http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.112.7819 _______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
_______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
On 10/1/2014 6:25 PM, Gershom B wrote:
It seems to me the far easier choice, in ditching the feature “not worth the problem” is simply to eliminate orphan instances as in “Extension 1” described in Edward Yang’s blog post (http://blog.ezyang.com/2014/09/open-type-families-are-not-modular/). While I definitely agree with you that orphan instances should go, or at least have some resolution mechanism, I'm proposing local instances on their own merits.
Gesh
participants (3)
-
Gershom B -
Gesh -
Jon Sterling